Motor neurons and oligodendrocytes arise from distinct ...genesdev.cshlp.org/content/29/23/2504.full.pdf · Motor neurons and oligodendrocytes arise from distinct cell lineages by

Motor neurons and oligodendrocytes arisefrom distinct cell lineages by progenitorrecruitmentAndrew M. Ravanelli and Bruce Appel

Department of Pediatrics, Department of Cell and Developmental Biology, University of Colorado School of Medicine, Aurora,Colorado 80045, USA

During spinal cord development, ventral neural progenitor cells that express the transcription factors Olig1 andOlig2, called pMN progenitors, produce motor neurons and then oligodendrocytes. Whether motor neurons andoligodendrocytes arise from common or distinct progenitors in vivo is not known. Using zebrafish, we found thatmotor neurons and oligodendrocytes are produced sequentially by distinct progenitors that have distinct origins.When olig2+ cells were tracked during the peak period of motor neuron formation, most differentiated as motorneurons without further cell division. Using time-lapse imaging, we found that, as motor neurons differentiated,more dorsally positioned neuroepithelial progenitors descended to the pMN domain and initiated olig2 expression.Inhibition of Hedgehog signaling duringmotor neuron differentiation blocked the ventral movement of progenitors,the progressive initiation of olig2 expression, and oligodendrocyte formation. We therefore propose that the motorneuron-to-oligodendrocyte switch results from Hedgehog-mediated recruitment of glial-fated progenitors to thepMN domain subsequent to neurogenesis.

[Keywords: oligodendrocytes; neural progenitors; motor neurons; zebrafish; Sonic hedgehog; pMN domain]

Supplemental material is available for this article.

Received September 15, 2015; revised version accepted October 30, 2015.

During vertebrate development, neural progenitor cellsdivide to expand the progenitor population and enlargethe neural tube and then differentiate as distinct typesof neurons and glial cells. One striking feature of neu-ral development is that neurons differentiate before glialcells. For example, pMN progenitors of the ventral spinalcord, defined by expression of the transcription factorOlig2, produce first motor neurons and then oligodendro-cytes, themyelinating glial cell type of the central nervoussystem (Fig. 1A). The mechanisms that maintain neuralprogenitors and guide the transition from neuron to glialcell production are key determinants of brain size andcomplexity.

Several lines of evidence indicate that the switch fromneuronal to glial cell production results from down-regu-lation of proneuronal factors and activation of proglial fac-tors. For example, spinal cord cells cease to express theproneuronal basic helix–loop–helix (bHLH) transcriptionfactors Ngn1 and Ngn2 prior to oligodendrocyte forma-tion (Zhou et al. 2001). This roughly coincides with theonset of Sox9 and NFIA/B expression, transcription fac-tors that promote gliogenesis (Stolt 2003; Deneen et al.

2006; Kang et al. 2012). Within pMN progenitors, revers-ible phosphorylation of Olig2 promotes the transitionfrom motor neuron to oligodendrocyte production, per-haps by regulating Olig2 interaction with Ngn2 (Li et al.2011). Signaling between neural cells, mediated by Notchreceptors, might also contribute to the neuron-to-gliaswitch because elevated Notch activity can block neuro-genesisandpromotegliogenesis (Gaianoetal.2000;Cham-bers et al. 2001; Park and Appel 2003). How these variousfunctions are integrated with the cell lineages that giverise to neurons and glia remains poorly understood.

In principle, individual progenitors could divide repeat-edly to sequentially generate neurons and glia. In this sce-nario, whichwe refer to as the common progenitormodel,the developmental potential of progenitors must changewith time to drive successive production of neurons andglia. Consistent with this possibility, lineage tracing tech-niques revealed that some neuroblasts of the fly nervoussystem function as common progenitors of neurons andglial cells (Bossing et al. 1997; Schmid et al. 1999), andindividual cells isolated from rodent embryonic neural

Corresponding authors: [email protected], [email protected] published online ahead of print. Article and publication date areonline at http://www.genesdev.org/cgi/doi/10.1101/gad.271312.115.

© 2015 Ravanelli and Appel This article is distributed exclusively byCold Spring Harbor Laboratory Press for the first six months after thefull-issue publication date (see http://genesdev.cshlp.org/site/misc/terms.xhtml). After six months, it is available under a Creative CommonsLicense (Attribution-NonCommercial 4.0 International), as described athttp://creativecommons.org/licenses/by-nc/4.0/.

2504 GENES & DEVELOPMENT 29:2504–2515 Published by Cold Spring Harbor Laboratory Press; ISSN 0890-9369/15; www.genesdev.org

Cold Spring Harbor Laboratory Press on May 1, 2020 - Published by genesdev.cshlp.orgDownloaded from

mailto:[email protected]







http://www.genesdev.org/cgi/doi/10.1101/gad.271312.115

http://www.genesdev.org/cgi/doi/10.1101/gad.271312.115

http://genesdev.cshlp.org/site/misc/terms.xhtml




http://creativecommons.org/licenses/by-nc/4.0/



http://genesdev.cshlp.org/

http://www.cshlpress.com

tissue and grown in culture also can produce both neu-rons and glia (Temple 1989; Reynolds et al. 1992; Davisand Temple 1994; Kalyani et al. 1997). Furthermore,time-lapse imaging revealed that neurons and glia are pro-duced in sequence in rodent cell culture (Qian et al. 2000),suggesting that mechanisms that trigger the neuron–glialswitch are intrinsic to individual cell lineages. Additionalsupport for the common progenitor model comes fromanalysis of the clonal progeny of single labeled neural pro-genitors in vivo. In particular, the progeny of some clonesmarked by recombinant retroviral infection in chick andfluorescent dye injection in zebrafish included bothmotorneurons and cells with the morphological characteristicsof oligodendrocyte lineage cells (Leber et al. 1990; Leberand Sanes 1995; Park et al. 2004).An alternative possibility is that neurons and glial cells

arise from distinct progenitors. In this case, which we

refer to as a heterogeneous progenitor model, distinct sub-sets of progenitors are restricted to either neuronal or glialfates. Consistent with this possibility, many clones aris-ing from single cultured neural cells isolated from embry-onic day 10.5–14 (E10.5–E14) mouse embryos includedonly neurons or only glia (Qian et al. 1998). Similarly,the majority of clonally related cell lineages in several ro-dent in vivo studies exclusively consisted of neurons orglial cells. For example, clones marked by retroviral infec-tion of ventricular zone cells in E12–E14 mouse embryosand E15–E16 rat embryos were almost entirely composedof only neurons or only glia (Luskin et al. 1988, 1993;Grove et al. 1993). Likewise, a significant proportion ofmouse telencephalon clones marked by retroviral infec-tion at E9.5 consisted of only glia (McCarthy et al.2001). The mechanisms that might specify fate-restrictedprogenitors and control the time at which they differenti-ate as neurons or glia are not known.Both common and distinct motor neuron and oligoden-

drocyte progenitors have been proposed based on gene ex-pression, gene function, and cell ablation data (Lu et al.2002; Zhou et al. 2002; Liu and Rao 2003; Rowitch 2004;Mukouyama et al. 2006; Wu et al. 2006). In this study,we aimed to resolve the uncertain lineage relationshipof motor neurons and oligodendrocyte precursor cells(OPCs) and identify the potential basis of neuronal and

Figure 1. Differentiation and division of pMNprogenitors are bi-ased by dorsoventral position. (A) Schematic diagramof the zebra-fish neural tube in transverse section depicting the ventraldomains (p0–p3) patterned along the Sonic hedgehog (Shh) signal-ing gradient. The pMN domain (green) is defined by expression ofolig2 and gives rise tomotor neurons (MN) and then oligodendro-cyte precursor cells (OPC) and oligodendrocytes (OL). (B) Sche-matic diagram of cell transplantation strategy. Cells weretransplanted between 1000-cell stage embryos. All transplantedcells express act2b:mCherry (red cells). Transplant-derivedpMN domain cells additionally express olig2:EGFP (green + redcells). Lateral views of chimeric embryos were imaged at 24 hpost-fertilization (hpf) at the level of the anus. (C ) Maximum in-tensity projection of confocal z-stack images of transplantedolig2:EGFP+ actb2:mCherry+ pMN progenitors. Panels showframes selected from a time-lapse movie (Supplemental MovieS1). The images show a lateral view of the trunk spinal cord,with dorsal to the top. Carets and arrows denote basal processesand axons, respectively. Over time, a basal process transformsinto an axon that extends posteriorly and ventrally. (D) Panelsshow frames selected from a time-lapse movie (SupplementalMovie S2). xy represents a single focal plane of a lateral view ofthe trunk spinal cord, with dorsal to the top. Orthogonal projec-tions of confocal z-stacks are denoted as xy and yz. xz representsa horizontal projection, and yz represents a transverse projection.Green cells reveal olig2:EGFP expression, and red nuclei resultfrom H2B-RFP expression. Dashed lines mark the central canal.An asterisk denotes an example of a single tracked cell. (E) Sche-matic representation of D xy t = 0. pMN domain cells appeargreen and are labeled based on their dorsoventral position withinthe pMN domain: dorsal (D), middle (M), or ventral (V). (F ) Quan-tification of olig2:EGFP+ pMN cell divisions from 24 to 42 hpf.n = 84 tracked nuclei from six embryos. (t) Time in minutesfrom the start of the movie. Bars, 10 µm.

Motor neuron and oligodendrocyte progenitors

GENES & DEVELOPMENT 2505




glial fate specification. To do so, we performed a series ofin vivo, time-lapse imaging experiments using zebrafishembryos, which allow direct visualization of neural pro-genitor behavior. Our data provide strong evidence thatat least the majority of motor neurons and OPCs arisefrom distinct progenitors. Importantly, progenitors thatproduce OPCs acquire pMN identity after those that pro-duce motor neurons instead of at the same time. By track-ing cell movements and gene expression, we learned thatneuroepithelial cells that originate dorsal to the pMNdomainmove ventrally and initiate olig2 expression, char-acteristic of pMN identity, concomitant with motorneuron differentiation. Inhibition of Hedgehog (Hh) sig-naling during motor neuron differentiation stalled ventralmovement of neuroepithelial cells and eliminated OPCdevelopment. We propose that ventral sliding of the neu-roepithelium brings new cells in range of Hh signals toreplenish pMN progenitors that differentiate as motorneurons. This process, which we call progenitor recruit-ment, provides a mechanistic basis for the sequentialproduction of motor neurons and OPCs from distinctprogenitors.

Results

pMN progenitor fate is biased by dorsoventral position

To test the fate of individual pMN progenitors, we firstperformed blastula stage transplantation followed by invivo confocal time-lapse microscopy. To do so, we trans-planted cells from olig2:EGFP;actb2:mCherry transgenicdonor embryos into stage-matched wild-type hosts (Fig.1B). In this experiment, all transplanted cells expressedmCherry from the actb2 (β-actin) reporter, which drivesubiquitous expression, and pMN progenitors and their de-scendents additionally expressed EGFP under control ofolig2 regulatory DNA (Shin et al. 2003). At 22–24 h post-fertilization (hpf), we sorted host embryos that had trans-planted EGFP+ pMN cells. Progenitors were readilyevident as cells having neuroepithelial morphologieswith apical membranes lining the ventricle and basal pro-jections connecting to the pial surface (Fig. 1C). We thenimaged individual EGFP+ mCherry+ pMN cells continu-ously until 42 hpf. Of 108 individual olig2:EGFP+ progen-itors, 90 differentiated into motor neurons withoutdividing (Fig. 1C; Supplemental Movie S1). Five cells re-mained as pMN progenitors without dividing, 11 cellsdivided to produce two new progenitors, and two cells di-vided once to produce two neurons. Notably, no cellsdivided asymmetrically to produce a neuron and a newprogenitor. These data lead us to three conclusions. First,many progenitors in the pMN domain at 24 hpf differenti-ate as motor neurons without dividing immediatelybeforehand. Second, pMN progenitor divisions that occurwithin this developmental period are symmetric, produc-ing either two new progenitors or two neurons. Third, thepMN progenitor pool does not appear to be maintained byasymmetric, self-renewing divisions.

In zebrafish, the pMN domain at 24 hpf is approximate-ly three cells deep on the dorsoventral axis. Is the probabil-

ity that a pMN progenitor differentiates as a neuron ordivides influenced by the dorsoventral position of a cellwithin the pMN domain? To gain insight into this ques-tion, we used time-lapse imaging to assess the behaviorsof individual progenitors within the pMN domain. Totrack individual cells, we injected cleavage stage olig2:EGFP embryos with mRNA encoding either DsRed2 fluo-rescent protein containing a nuclear localization motif(dsRed2nuc) or the human histone protein HIST2H2BEfused to RFP (H2B-RFP).We collected confocal z-stack im-ages at the level of the trunk spinal cord from 24 to 41 hpf.For this experiment, we defined pMNprogenitors as olig2:EGFP+ cells with nuclei positioned adjacent to the spinalcord lumen (Fig.1D).Asexpected,pMNprogenitorsmovedradially from the spinal cord lumen and differentiated asmotorneurons (Fig.1D;SupplementalMovieS2);however,we found a clear bias based on starting position. Whereasall ventral pMN progenitors produced motor neurons(n = 38), only 71% (n = 14) and 22% (n = 32) of middle anddorsal pMN progenitors, respectively, generated motorneurons.

By tracking mitoses, we found that progenitor position(Fig. 1E) also correlated with cell division. All ventral(n = 38) and 12 of 14 middle pMN progenitors differentiat-ed asmotor neuronswithout dividing (Fig. 1F). In contrast,themajority of dorsal pMNprogenitors divided one to twotimes (n = 32) (Fig. 1F). Most divisions produced two newpMN progenitor progeny that remained associated withthe lumen and did not differentiate during the imaging pe-riod. These data indicate that ventral pMN progenitors di-vide rarely and differentiate as motor neurons, whereasmore dorsal pMN progenitors tend to remain proliferativeand undifferentiated. Thus, whether a pMN progenitordifferentiates as a neuron or divides correlates with its po-sition on the dorsoventral axis, raising the possibility thatspatial cues operate within the pMN domain to specifyprogenitor fate.

Motor neurons andOPCs arise from distinct cell lineagesthat initiate olig2 expression at different times

Our data indicate that, at 24 hpf, most olig2:EGFP+ pMNprogenitors that differentiate asmotor neurons do sowith-out dividing. This observation is inconsistent with a com-mon lineage model in which motor neurons and OPCsdevelop from a common progenitor and instead suggeststhat they arise from different cell lineages (Fig. 2A). What,then, is the origin of progenitors that produce OPCs? Onepossibility, whichwe refer to as a heterogenous progenitormodel, is that olig2:EGFP+ pMN progenitors are mixedand fate-restricted, whereby some produce only motorneurons and others, perhaps those in dorsal pMN, produceonly OPCs. A second possibility, which we call a pro-genitor recruitment model, is that pMN progenitors thatdevelop as OPCs initiate olig2:EGFP expression afterthose that develop as neurons. To begin to distinguish be-tween these possibilities, weused photoconvertible Kaedefluorescent protein expressed by an olig2 transgene (Zan-nino and Appel 2009) to determine whether cells thatdevelop as neurons and OPCs initiate olig2 expression at

Ravanelli and Appel

2506 GENES & DEVELOPMENT




the same or different times. When exposed to ultraviolet(UV) light, green Kaede (olig2:Kaedegreen) is converted tored (olig2:Kaedered). Thus, at the time of photoconversion,all olig2:Kaedegreen-expressing progenitors turn red, andtheir progeny inherit the redKaedeprotein.Motorneuronsdown-regulate olig2 expression; therefore, most motorneurons derived from photoconverted progenitors remainred because they no longer express new olig2:Kaedegreen.pMNcells andOPCsderived fromphotoconvertedprogen-itors continue to express olig2:Kaedegreen and appear yel-low (Kaedeyellow) from a combination of photoconvertedand newly synthesized unconverted Kaede. Thus, fromboth the common and heterogeneous progenitor models,we predict that photoconversion during the period ofmotor neuron differentiation would result in red motorneurons and yellowOPCs (Fig. 2B). In contrast, the progen-itor recruitment model leads to the prediction that photo-conversion during motor neuron differentiation wouldresult in redmotor neurons and greenOPCs, if progenitorsthat produce OPCs initiate olig2:Kaede expression afterthose that produce motor neurons (Fig. 2B).To test these predictions, we photoconverted embryos

at 24 hpf and analyzed them at 72 hpf. Motor neurons

that migrated dorsolaterally were red as a consequenceof the photoconverted Kaede (Fig. 2C–E). In contrast,dorsally migrated OPCs as well as numerous cells with-in the pMN domain were green, indicating that they ini-tiated olig2 expression after photoconversion at 24 hpf(Fig. 2C–E). To determine when oligodendrocyte lineagecells initiated olig2:Kaedegreen expression, we photocon-verted Kaede in different groups of embryos at 2-h inter-vals. This revealed a progressive decrease in the numberof OPCs that were only green and a concomitant in-crease in the number of yellow OPCs (Fig. 2F–I). Similar-ly, there appeared to be more olig2:Kaedeyellow cells andfewer olig2:Kaedegreen cells within the pMN domainwith progressively later photoconversions (Fig. 2F–I).These data support two main conclusions. First, becausecells that produce motor neurons and OPCs initiateolig2 expression at different times, motor neurons andOPCs must arise from distinct pMN progenitors. Sec-ond, the sequential expression of olig2 by OPCs andsome pMN cells following motor neuron formation indi-cates that ventral spinal cord cells acquire pMN identityprogressively, consistent with the progenitor recruit-ment model.

Figure 2. Motor neurons and oligodendro-cytes arise from distinct cell lineages thatprogressively initiate olig2 expression. (A)Two contrasting models of pMN lineages.Under thecommonprogenitormodel, an in-dividual pMN progenitor can produce bothmotor neurons (MN) and oligodendrocytes(OL). Under the heterogeneous progenitormodel, motor neurons and oligodendro-cytes arise from distinct, lineage-restrictedpMN progenitors. (B) Schematic of olig2:Kaede photoconversion strategy and experi-mental predictions.Red cells represent pho-toconverted Kaede (Kaedered). Green cellsrepresent those that initiate Kaedegreen ex-pression after photoconversion. Yellowcells represent those that containKaedegreen

and Kaedered (Kaedeyellow). White cells(olig2−) represent those that do not expressolig2:Kaede at the time of photoconversion.The purple bolt indicates the time of UVlight illumination for photoconversion. (C )Lateral viewof a 72-hpfolig2:Kaede embryoin which Kaede was photoconverted at24 hpf. Asterisks mark Kaedered motor neu-rons, and arrows indicate Kaedegreen OPCs.Kaedeyellow cells may appear orange dueto the variable fluorescence intensity ofKaedered to Kaedegreen. (D) Transverse (yz)view, dorsal to the top, of a confocal z-pro-jection of a 72-hpf olig2:Kaede embryo pho-

toconverted at 24 hpf. The arrowdenotes aKaedegreenOPC. (E) Schematic representation of the image shown inD. Asterisksmark putativepMN progenitor cells. The dashed line denotes the central canal. (F ) Lateral view of 72-hpf olig2:Kaede embryo photoconverted at 36 hpf.Asterisks mark Kaedered motor neurons, the arrow indicates a Kaedegreen OPC, and arrowheads mark OPCs with both Kaedegreen andKaedered. (G) Transverse (yz) view from confocal z-projection of 72-hpf olig2:Kaede embryo photoconverted at 36 hpf. (H) Schematic rep-resentation of the image shown inG. (I ) Graph showing the percentage of total dorsallymigratedOPCs that were Kaedeyellow or Kaedegreen

(n = 3136 OPCs from 114 embryos) in olig2:Kaede embryos photoconverted at the indicated time points. Error bars indicate SEM. Bars10 µm.






Ventral movement of neuroepithelial cells contributesto pMN progenitor recruitment

One possible explanation of our data is that fate-restrictedprogenitors within the pMN domain initiate olig2 expres-sion at different times. However, as in mice (Mukouyamaet al. 2006), all cells within the pMN domain appear to ex-press olig2. An alternative explanation is that new cellsmove into the pMN domain and initiate olig2 expressionconcomitant with motor neuron formation. If so, theappearance of olig2:Kaedegreen following photoconversionmight reveal the source of new progenitors recruited tothe pMN domain. We therefore photoconverted olig2:Kaede at 24 hpf and performed time-lapse imaging to as-sess nascent olig2:Kaedegreen expression. During the next13 h, numerous olig2:Kaedegreen cells appeared almost ex-clusively at the dorsal edge of the previously photocon-verted olig2:Kaedered cells (Fig. 3A; Supplemental MovieS3). Thus, cells dorsal to the original olig2+ pMN domainadopt pMN identity during the period of motor neurondifferentiation.

One potential reason for why cells close to the dorsalboundary of the pMN domain initiate olig2 expression isthat the pMN domain expands dorsally to include morecells. Alternatively, more dorsally positioned cells mightmove ventrally to enter the pMN domain. To distinguishbetween these possibilities, we expanded our initial analy-sis of the tracking data described in Figure 1 to follow cellsoutside the pMN domain. This analysis showed that cellslocated within one to two cell diameters of the dorsaledge of the pMNdomain descended ventrally and initiatedolig2:EGFPexpression (Fig. 3B,C;SupplementalMovieS4).At the same time, cells already within the pMN domainshifted ventrally and differentiated as they reached moreventral positions (Fig. 3B; Supplemental Movie S5). Thesedata indicate that, as motor neurons differentiate, newpMN progenitors are recruited from more dorsally posi-tioned cells that move ventrally to enter the domain.

To helpmap the origin of cells that descend to the pMNdomain and investigatewhether ventralmovement of pro-genitors is a general feature of spinal cord development,we determined the distribution of cells that arise fromp0 domain progenitors. To do so, we used the transgenesdbx1b:GFP (Satou et al. 2012) to transiently mark p0domain cells and dbx1b:Cre (Satou et al. 2012) to mediaterecombination of a ubi:Zebrabow reporter (Pan et al.2013), therebyproviding apermanentmarkerof p0domainprogeny. At 72 hpf, dbx1b:GFP expression was limited tocells dorsal to those that expressed Nkx6.1, which markscells of the p2–p3 domains (Fig. 3D,E; Briscoe and Ericson1999). In contrast, someNkx6.1+ cellsweremarkedbyubi:Zebrabow expression (Fig. 3F,G). We interpret these re-sults to mean that some cells that originated in p0 movedventrally to the p2 domain, terminated dbx1b expressionas they exited the p0 domain, and initiated nkx6.1 expres-sion as they entered the p2 domain. These observationssupport several conclusions. First, ventral movement ofprogenitors is not limited to those that descend into thepMN domain but also includes more dorsal cells, raisingthe possibility that similar progenitor movements occur

throughout the spinal cord. Second, p0domainprogenitorsdo not enter the pMN domain, indicating that the likelyorigin of newly recruited pMNprogenitors is the p1–p2 do-mains. Third, movement of progenitors between domainsand concomitant changes in the gene expression profilescharacteristic of those domains indicate that progenitorsretain fate plasticity.

Ventrally moving cells retain neuroepithelial features

Do spinal cord progenitors delaminate from the neuroepi-thelium and move as individuals or retain their neuroepi-thelial characteristics during ventral movement betweenprogenitor domains? To distinguish between these possi-bilities, we injected cleavage stage olig2:Kaede embryoswith mRNA encoding a functional fusion of the apicalPar polarity protein Pard3 toGFP (Tawk et al. 2007). Theseembryoswere repeatedlyexposed to 405-nm light tomain-tain Kaedered photoconversion, allowing visualization ofPard3-GFP. At 24 hpf, we identified Kaede− Pard3-GFP+

neuroepithelial cells positioned dorsally to the pMN do-main and tracked their movements using time-lapse con-focal microscopy. Tracked cells that moved ventrally andsubsequently initiated olig2:Kaede expression retainedapical localization of Pard3-GFP, indicating that they re-mained as neuropethelial cells (Fig. 4A; SupplementalMovie S6). Cells retained apical localization of Pard3-GFP even during cell division and only down-regulatedPard3-GFP concomitant with lateral movement of thesoma during differentiation (Fig. 4B; Supplemental MovieS7). These data indicate that progenitors recruited to thepMN domain preserve neuroepithelial characteristics, in-dicating that the neuroepithelium shifts ventrally as ven-tral pMN progenitors delaminate, move radially, anddifferentiate as neurons.

pMN progenitor recruitment requires Hh signaling

Experiments using explanted neural tube tissue showedthat similar concentrations of Shh induced formation ofboth motor neurons and oligodendrocytes (Pringle et al.1996; Orentas et al. 1999). How these distinct cell typesare specified by the same level of Shh signaling is notknown. Our observation that some dorsally positionedprogenitors acquire pMN identity led us to hypothesizethat ventral cell movement repositions progenitors with-in the Shh morphogen gradient, causing them to initiateolig2. If so, blocking Shh signaling at the time of motorneuron formation should interfere with pMN progenitorrecruitment, resulting in fewer olig2+ cells and a deficitof OPCs. To test this prediction, we treated olig2:EGFPembryos injected at cleavage stage with mRNA encodingdsRed2nuc with cyclopamine, which effectively reducesHh signal transduction, at 24 hpf. Cell tracking revealedthat cells within one to two cell diameters of the dorsaledge of the pMN domain did not move ventrally and didnot express olig2:EGFP (Fig. 5A; Supplemental MovieS8). Additionally, there were fewer EGFP+ cells at theend of the imaging period in cyclopamine-treated embryos(average of 31.7 ± 1.9 EGFP+ cells per hemisegment; n = 4

Ravanelli and Appel





Figure 3. The pMNdomain ismaintained by recruitment of new progenitors thatmove ventrally within the spinal cord. (A) Panels showframes selected from a time-lapse movie (Supplemental Movie S3). (t) Time in minutes from the start of the movie. Images show lateralviews of the trunk spinal cord with dorsal to the top. (Top panels) At 24 hpf, olig2:Kaede was fully photoconverted from green to red andimaged continuously. Dashed lines mark the most dorsal olig2:Kaedered ventricular cells. (Middle panels) By 7.5 h post-conversion(29.5 hpf), faint olig2:Kaedegreen is evident in some cells (green arrows) dorsal to the pMN boundary. (Bottom panels) By 11 h post-conver-sion (35 hpf), numerous cells at the dorsal edge of the original pMNboundary areKaedegreen. Red arrowsmarkKaederedmotor neurons thathave migrated dorsolaterally. Yellow arrowsmark Kaedeyellow pMN cells that continue to express olig2:Kaede. Asterisks denote dorsolat-erally migrated olig2:Kaedered neurons out of the plane of focus. (B) Panels show frames selected from a time-lapse movie (SupplementalMovies S4, S5) of an olig2:EGFP embryo injected with mRNA encoding dsRed2nuc. Images are focused on the trunk spinal cord from lat-eral orientation, and dorsal is to the top. Nuclei of EGFP− (o) and EGFP+ (+) cells were tracked continuously, beginning at 24 hpf. Thedashed line represents the dorsal border of the EGFP+ pMN domain. By 39 hpf, numerous cells that originated dorsal to the pMN domainoccupied positions within the pMN domain and expressed olig2:EGFP. The top panels showmerged images, and the bottom panels showonly EGFP expression. (C ) Graph showing the percentage of tracked EGFP− cells at various distances from the dorsal border of the pMNdomain that initiated expression of olig2:EGFP. n = 218 cells from 15 embryos. (D) Transverse section of a 72-hpf dbx1b:GFP zebrafishtrunk spinal cord. White arrowheads denote a dbx1b:GFP+ p0 progenitor. Brackets denote p2–p3 domains marked by Nkx6.1 expression.(E) Schematic representation of Nkx6.1+ progenitors (red) and GFP+ cells (green). (F ) Transverse section of a 72-hpf zebrafish trunk spinalcord. dbx1b:Cre-converted ubi:Zebrabow cells were stained with anti-GFP antibody (green). p2–p3 domain progenitors were stained withanti-Nkx6.1 antibody (red). The arrow indicates an Nkx6.1+ p2 cell that expressed dbx1b:Cre. (G) Schematic representation of Nkx6.1+

progenitors (red) and Cre-converted cells (green). The double-positive cell in p2 domain appears red/green. Bars, 10 µm.






embryos) than in control embryos (average of 38.2 ± 0.3EGFP+ cells per hemisegment; n = 4 embryos; P < 0.01). Fi-nally, and consistent with our previous data showing thatcontinuous Hh signaling is required for OPC formation(Park et al. 2004), embryos treated at 24 hpf with cyclop-

amine had fewer OPCs at 72 hpf than controls, in adose-dependent manner (Fig. 5B–D). We interpret thesedata to mean that ventral movement maintains the pMNdomain by bringing progenitors within range of Hh signalsthat specify pMN identity and that progenitor recruit-ment by continuous Hh signaling is necessary for OPCdevelopment.

Discussion

In vertebrate embryos, neural cell fate is determined inpart by positional information. For example, the dorsoven-tral axis of the neural tube is patterned by graded distribu-tion of themorphogens Shh and Bmp, resulting in distinctprogenitor populations that produce different kinds ofneurons. Importantly, spatial patterning is integratedwith temporal patterning to diversify neural cell fate. Inparticular, after producing neurons, progenitors switch toglial cell production. These features of neural develop-ment raise a fundamental question: How do progenitorsthat are exposed to a particular combination of spatialcues give rise to different cells over time? Here we pro-vide evidence that, as motor neurons differentiate, devel-opmentally plastic progenitors are repositioned withinthe Hh signaling gradient to be specified for oligodendro-cyte fate.

Onemechanism bywhich neural progenitors could pro-duce different cells at different times is by progressingthrough a series of asymmetric, self-renewing divisions.A good example of this is provided by Drosophila neuro-blasts, which, as they divide, express a sequence of tran-scription factors that specify progeny cells for distinctneuronal fates (Isshiki et al. 2001; Cleary 2006; Kaoet al. 2012; Bayraktar and Doe 2013). Whether a similartype of mechanism specifies the progeny of self-renewingpMN progenitors for motor neuron and oligodendrocytefates has not been clear. Neural explant and gene expres-sion studies suggested that many oligodendrocytesdevelop from the ventral spinal cord, close to motor neu-rons (Warf et al. 1991; Noll and Miller 1993; Pringle andRichardson 1993; Yu et al. 1994; Timsit et al. 1995), andare therefore consistent with the possibility of a commonprogenitor origin. Also consistent with a common progen-itor hypothesis are observations that both motor neuronand oligodendrocyte progenitors express the Olig1 andOlig2 transcription factors and that both cell types failto develop in mouse and zebrafish embryos that lackOlig functions (Lu et al. 2002; Park et al. 2002; Takebaya-shi et al. 2002; Zhou et al. 2002). Furthermore, analysis ofthe clonal progeny of neural cells marked by viral or fluo-rescent tracers in chicks and zebrafish indicated that mo-tor neurons and oligodendrocytes can arise from commonlineages (Leber et al. 1990; Leber and Sanes 1995; Parket al. 2004). However, most of the reported clones con-tained only neurons, suggesting that a shared lineage isnot obligatory for motor neuron and oligodendrocytedevelopment. Indeed, more recent studies revealed thatdorsal progenitors that are far removed from motor neu-rons also produce oligodendrocytes (Cai et al. 2005;

Figure 4. Progenitor cells retain neuroepithelial characteristicswhile they move ventrally to the pMN domain. Panels showframes selected from time-lapse movies. (t) Time in minutesfrom the start of the movie. olig2:Kaede embryos were injectedwith CMV:pard3-GFP DNA and photoconverted repeatedly dur-ing imaging to allow visualization of Pard3-GFP (green). Kaedered

cells appearwhite. Orthogonal (yz [A,B] and xy [A′,B′]) projectionsof confocal z-stacks of a trunk spinal cord imaged from 24 hpf.Dashed white lines denote the central canal. The dotted whiteline outlines the ventral neural tube. (A) Arrowheads denotePard3-GFP apical localization in olig2:Kaedered-negative cells at24 hpf that move ventrally over time to become olig2:Kaedered-positive (Supplemental Movie S6). (A′ ′) Insets from the t = 1075-min transverse (yz) image are contrast-enhanced to show coex-pression of olig2:Kaedered and Pard3-GFP. The tracked cell is out-lined by a dotted magenta line. Arrows denote cell divisionwithin the neuroepithelium in successive time points. (B)Pard3-GFP apical localization in olig2:Kaedered-positive progeni-tors is lost upon differentiation and dorsolateral migration awayfrom the central canal (denoted by dashed white lines) (Supple-mental Movie S7). The top panels show merged images (olig2:Kaedered with Pard3-GFP), and the bottom panels show onlyPard3-GFP. The tracked cell is outlined by a dotted magentaline. Bars: A,B, 10 µm; A′′, 5 µm.

Ravanelli and Appel





Fogarty et al. 2005; Vallstedt et al. 2005), showing thatmotor neuron and oligodendrocyte specification can beuncoupled. In this study, we present two new lines ofevidence that at least the majority of motor neuronsand oligodendrocytes arise from distinct pMN progeni-tor cell lineages. First, our time-lapse imaging showedthat most motor neurons differentiated from pMN pro-genitors without undergoing a division to produce anoligodendrocyte-fated progenitor. Second, our Kaede pho-toconversion experiments showed that motor neuron andoligodendrocyte progenitors initiated olig2 expression atdifferent times, which is most compatible with the ideathat they belong to different lineages. We therefore con-clude that most motor neuron and oligodendrocytes do

not arise from shared lineages but instead from distinctprogenitors.If distinct pMN progenitor cells produce motor neurons

and oligodendrocytes, how are they specified? One possi-bility is that pMN progenitors have equivalent potentialfor motor neuron and oligodendrocyte fates but are sto-chastically selected to develop as one or the other celltype. Consistent with this possibility are observationsthat Notch signaling, which stochastically selects ecto-dermal cells for neuroblast fate in Drosophila (Simpson1997), limits motor neuron formation and promotesOPC development in zebrafish (Park and Appel 2003).However, neural cell lineages were not observed directlyin embryos having deficient or hyperactive Notch signal-ing, leaving the relationship of distinct pMN progenitorsto one another and the basis of Notch-mediated fate spec-ification unclear. Recent data showing that Notch activi-ty increases the sensitivity of neural cells to Shh (Huanget al. 2012; Kong et al. 2015; Stasiulewicz et al. 2015) raisethe possibility that Notch regulation of motor neuron andoligodendrocyte fates results from modulation of Hh sig-naling instead of by mediating binary fate decisionsamong equivalent progenitors. Our time-lapse imagingstudies now reveal how distinct progenitors mightrespond to discrete levels of Notch and Hh signaling ac-tivity over time to form motor neurons and oligodendro-cytes. By tracking cell division and fate, we uncoveredevidence of a differentiation gradient within the pMNdomain in which the ventral-most pMN progenitors dif-ferentiate as motor neurons without dividing, whilemore dorsal pMN cells continue to divide. Surprisingly,in a process that we call progenitor recruitment (Fig. 6),we also found that neuroepithelial progenitors that origi-nated dorsal to the pMN domain moved ventrally and ini-tiated olig2 expression at the same time that ventral pMNprogenitors moved radially and differentiated as motorneurons. Cells added to the pMN domain likely includeoligodendrocyte-fated progenitors because they initiateolig2 expression after motor neuron-fated progenitors.Addition of new progenitors to the pMN domain requiresHh signaling because blocking Hh eliminated ventral cellmovement, reduced the number of pMN cells, and abol-ished OPC formation. We therefore propose that progeni-tor recruitment repositions cells within the Hh signalinggradient, leading to the sequential production of motorneurons and oligodendrocytes (Fig. 6). Importantly, ourmodel provides the basis for a new interpretation of thesubset of clones in our previous work that containedboth motor neurons and oligodendrocytes (Park et al.2004). We now speculate that mixed clones arose frommore dorsal, highly proliferative progenitors that pro-duced progeny cells that moved ventrally and initiatedolig2 expression at different times to become motor neu-rons and OPCs.In mouse and chick embryos, different spinal cord pro-

genitor domains grow to distinct sizes because the cellswithin them differentiate at dissimilar rates (Kichevaet al. 2014). In particular, the differentiation rate of pMNprogenitors peaks at the time of motor neuron develop-ment, preceding the differentiation peak of more dorsal

Figure 5. Progenitor recruitment and OPC specification requireHh signaling. (A) Panels show frames selected from a time-lapsemovie (Supplemental Movie S8) of an olig2:EGFP embryo inject-ed with mRNA encoding H2B-RFP. Images are focused on thetrunk spinal cord from lateral orientation, and dorsal is to thetop. Embryoswere treatedwith 5 µMcyclopamine at 24 hpf to in-hibit Hh signaling and were imaged continuously. Nuclei ofEGFP− (∗) cells were tracked, beginning at 24 hpf. The dashedline represents the dorsal border of the EGFP+ pMN domain. By40 hpf, cells that originated dorsal to the pMN domain remainedat their original dorsoventral position and did not initiate expres-sion of olig2:EGFP. The top panels showmerged images, and thebottom panels show only EGFP expression. (B,C ) Maximum in-tensity projection of confocal z-stack images of olig2:EGFP em-bryos treated with vehicle control or 5 µM cyclopamine at 24hpf and imaged at 72 hpf. Images are focused on the trunk spinalcord with dorsal to the top. Dorsally migrated OPCs are markedwith white arrows. The pMN domain is marked with brackets.(D) Graph showing the number of dorsally migrated OPCs perbilateral hemisegment in control (n = 10 embryos), 1 µM cyclop-amine-treated (n = 10 embryos), or 5 µM cyclopamine-treated(n= 11 embryos) embryos . (∗) P < 0.001. Error bars indicate SEM.Bars, 10 µm.






progenitor domains (Kicheva et al. 2014). Could progenitorrecruitment from more dorsal populations maintain thepMN domains of rodents and birds past the peak of differ-entiation? Individual neural cells and neural explants canchange domain identity if culture conditions are altered(Gabay et al. 2003; Dessaud et al. 2010), consistent with arequirement that recruited progenitors change domainidentity as they move ventrally. In mice, some progeny ofDbx1-expressing progenitors occupied the pMN domainand expressed Olig2 (Dessaud et al. 2010). This result wasinterpreted to mean that ventral spinal cord progenitorstransiently expressDbx1. Our results now raise the possi-bility that, instead, progenitors that originate within theDbx1+ p0 domain of mice move ventrally to contributeto the pMN domain. In mice programmed to express acell-lethal toxin in Olig1:Cre+ pMN progenitors, Olig2+

progenitors were maintained despite loss of most motorneurons and all OPCs (Wu et al. 2006). The investigatorsof this studyspeculated thatOlig2+progenitorswereregen-erated by a residentOlig2− pMNprogenitor and proposed asequential specificationmodel similar tomodels proposedfor Drosophila neuroblasts (Isshiki et al. 2001) and retinalprogenitors (Cepko et al. 1996). However, few, if any,Olig2− cells occupy the pMN domain (Mukouyama et al.2006). Our data now raise the possibility that Olig2+ pro-genitors are replaced following pMN cell ablation by re-

cruitment of more dorsal progenitors that initiate Olig2expression following ventral movement.

A basic requirement of progenitor cell populations isthat progenitors both are maintained to prevent prema-ture depletion and differentiate as distinct cell typeson schedule. This requirement can be filled by mecha-nisms that guide progenitors through asymmetric, self-renewing divisions. Our results now indicate that notall progenitor populations are self-sustaining. Instead,populations that undergo peak differentiation early inneural development can be maintained by recruitingnearby progenitors. Progenitor recruitment exposes pro-genitors to similar instructive signals at different times,thereby providing a mechanistic basis for the neuron-to-glia fate switch characteristic of vertebrate neuraldevelopment.

Materials and methods

Zebrafish strains and husbandry

The University of Colorado Anschutz Medical Campus Institu-tional Animal Care and Use Committee approved all zebrafishstudies. Embryos were raised at 28.5°C in E3 medium (5 mMNaCl, 0.17 mM KCl, 0.33 mM CaCl, 0.33 mM MgSO4 atpH 7.4, with sodium bicarbonate) and staged according to hourspost-fertilization and morphological criteria (Kimmel et al.1995). Transgenic alleles included Tg(olig2:eGFP)vu12 (Shinet al. 2003), Tg(olig2:Kaede)vu85 (Zannino and Appel 2009), Tg(dbx1b:Cre)nns13a (Satou et al. 2012), Tg(ubi:Zebrabow)a132Tg

(Pan et al. 2013), and Tg(dbx1b:GFP)nns11Tg (Satou et al. 2012).

mRNA synthesis and microinjection

mRNAs for dsRed2nuc (Clontech) and H2B-RFP (gift from AlexSchier) were synthesized with the mMessage mMachine kit(Ambion) using the manufacturer’s protocol and using linearizedplasmids as templates. mRNAswere injected at 250 ng/µL in a fi-nal solution of 0.2 M KCl with phenol red at the one-cell stage, at1- to 3-nL volumes.

Time-lapse imaging and cell tracking

Confocal time-lapse imaging was carried out by embedding em-bryos in 0.6% low-melting-point agarose (IBI Scientific) with6% tricaine (Sigma) and viewing the spinal cord in the transverseplanes using a 20× (NA 0.8), 40× oil immersion (NA 1.3), or 40×long-working-distance water immersion (NA 1.1) objective. Em-bryos were maintained at 28.5°C in a heated stage chamber, andz-stacks were collected at variable intervals, usually starting at24 hpf and continuing through to 48 hpf. Imageswere taken usinga Zeiss Axiovert 200 microscope equipped with a spinning diskconfocal system (PerkinElmer Improvision). Image brightnessand contrast were adjusted in Volocity (PerkinElmer) or ImageJ(National Institutes of Health). For comparison purposes, thesame adjustment was applied to all images. Time-lapse videoswere exported from Volocity as extended z-projection TIFF imag-es.We used ImageJ to rotate, crop, and translate in order to correctfor x–y drift. Image stacks were then exported in QuickTime(.MOV) format. Individual cells were tracked by following nucleilabeled with either H2B-RFP or dsRed2nuc at each time point af-ter image acquisition. Transgene expression, dorsoventral

Figure 6. Progenitor recruitment model. Individual panels por-tray changes in the ventral spinal cord during motor neuron andOPC specification. Initially, a Shh morphogen gradient patternsthe ventral neural tube to form the ventral progenitor domains.olig2+ pMN (green) and olig2− p2 and p3 domains (white) are de-picted. (Panels 1–5) Next, the most ventral pMN progenitors,schematically represented by cell A followed by cell B, differenti-ate as motor neurons without dividing. (Panels 2–6) As motorneurons evacuate the ventral spinal cord, more dorsally posi-tioned neuroepithelial progenitors, represented by cell C followedby cell D, move ventrally within the Shh gradient and initiateolig2 expression. (Panel 7) Recruited progenitors give rise toOPCs.

Ravanelli and Appel





position, and cell morphology were noted for each cell at the be-ginning and end of each image series.

Immunohistochemistry

Embryosweremanually dechorionated and fixed overnight in 4%paraformaldehyde in PBST (0.1% Tween-20). Twenty-microme-ter cryosections were used for immunohistochemistry with rab-bit anti-GFP (1:100; Invitrogen, A11122), mouse anti-Nkx6.1(1:500; Developmental Studies Hybridoma Bank, F55A12), goatanti-rabbit Alexa fluor488 (1:200; Invitrogen), or goat anti-mouseAlexa fluor568 (1:200; Invitrogen). ubi:Zebrabow embryos wereprocessed for immunohistochemistry following antigen retrieval(20 min at 98°C in citrate buffer [10 mM sodium citrate, 0.05%tween-20 at pH 6.0]) to eliminate broad-spectrum fluorescenceproduced by the transgenic prior to antibody labeling.

Kaede photoconversion

olig2:Kaede transgenic embryos were illuminated with UV lightfrom a mercury bulb source for 30–120 sec following manualdechorionation or for 10 sec with a 405-nm laser (20% power)on a spinning disk confocal microscope. Immediately afterward,photoconversion was confirmed by observation of the presenceof red fluorescence and an absence of green fluorescence. Photo-converted embryos were then kept at 28.5°C in the dark untilimaging.

Quantitative and statistical methods

Quantificationswere performed by collecting confocal z-stacks ofentire embryos and counting cells from maximally projected im-ages. Z images were examined to distinguish between individualcells. We plotted all data and performed all statistical analyses inMicrosoft Excel. All data are expressed asmean ± SEM. For statis-tical analysis, we used Student’s two-tailed t-test for all data withnormal distributions. Unless otherwise noted, statistical sig-nificance is indicated as follows: P < 0.05 (∗), P < 0.01 (∗∗), andP < 0.001 (∗∗∗).

Drug treatments

For Smoothened inhibitor assays, cyclopamine (Cayman Chemi-cal, 11321) was reconstituted in ethanol to make a 10 mM stockand stored at −20°C. Embryos were treated at 24 hpf with 1 or5 µM cyclopamine or an equal volume of ethanol alone for avehicle control in E3 medium.

Acknowledgments

We thank members of the Appel laboratory for valuable discus-sions, Santos Franco for comments on themanuscript, and Chris-tinaKearns for helpwith sectioning and antibody staining.We aregrateful for the gifts of the H2B-RFP plasmid and the Tg(ubi:Zebrabow)a132Tg line from Alex Schier, the CMV:Pard3-GFPplasmid from Jon Clarke, and the actb2:mCherry constructfrom Ken Poss. The mouse anti-Nkx6.1 (1:500; Develop-mental Studies Hybridoma Bank F55A12), developed by OleD. Madsen, was obtained from the Developmental Studies Hy-bridoma Bank, created by the National Institute of Child Healthand Human Development of the National Institutes of Healthandmaintained at TheUniversity of IowaDepartment of Biology.This work was supported by National Institutes of Health grantR01 NS40660 and a gift to B.A. from the Gates Frontiers Fund.

A.M.R. was supported by the National Institutes of Health (Na-tional Cancer Institute) T32 5T32CA08208613 Fellowship. TheUniversity of Colorado Anschutz Medical Campus ZebrafishCore Facility was supported by National Institutes of Healthgrant P30 NS048154. A.M.R. performed all experiments and con-ducted analyses of the data. A.M.R. and B.A. conceived the pro-ject and wrote the manuscript.

References

Bayraktar OA,DoeCQ. 2013. Combinatorial temporal patterningin progenitors expands neural diversity. Nature 498: 449–55.

Bossing T, Udolph G, Doe CQ, Technau GM, Schmidt H, RickertC, Bossing T, Vef O, Urban J, Technau GM. 1997. The embry-onic central nervous system lineages of Drosophila mela-nogaster. II. Neuroblast lineages derived from the dorsal partof the neuroectoderm. Dev Biol 189: 186–204.

Briscoe J, Ericson J. 1999. The specification of neuronal identityby graded Sonic hedgehog signalling. Semin Cell Dev Biol10: 353–362.

Cai J, Qi Y, Hu X, Tan M, Liu Z, Zhang J, Li Q, Sander M, Qiu M.2005. Generation of oligodendrocyte precursor cells frommouse dorsal spinal cord independent of Nkx6 regulationand Shh signaling. Neuron 45: 41–53.

Cepko CL, Austin CP, Yang X, Alexiades M, Ezzeddine D. 1996.Cell fate determination in the vertebrate retina. Proc NatlAcad Sci 93: 589–595.

Chambers CB, Peng Y, Nguyen H, Gaiano N, Fishell G, Nye JS.2001. Spatiotemporal selectivity of response to Notch1 sig-nals in mammalian forebrain precursors. Development 128:689–702.

ClearyMD. 2006. Regulation of neuroblast competence:multipletemporal identity factors specify distinct neuronal fates with-in a single early competencewindow.GenesDev 20: 429–434.

Davis A, Temple S. 1994. A self-renewing multipotential stemcell in embryonic rat cerebral cortex. Nature 372: 263–266.

Deneen B, Ho R, Lukaszewicz A, HochstimCJ, Gronostajski RM,Anderson DJ. 2006. The transcription factor NFIA controlsthe onset of gliogenesis in the developing spinal cord.Neuron52: 953–968.

Dessaud E, Ribes V, Balaskas N, Yang LL, Pierani A, Kicheva A,Novitch BG, Briscoe J, Sasai N. 2010. Dynamic assignmentand maintenance of positional identity in the ventral neuraltube by the morphogen Sonic hedgehog. PLoS Biol 8:e1000382.

Fogarty M, Richardson WD, Kessaris N. 2005. A subset of oligo-dendrocytes generated from radial glia in the dorsal spinalcord. Development 132: 1951–1959.

Gabay L, Lowell S, Rubin LL, Anderson DJ. 2003. Deregulation ofdorsoventral patterning by FGF confers trilineage differentia-tion capacity on CNS stem cells in vitro.Neuron 40: 485–499.

GaianoN,Nye JS, Fishell G. 2000. Radial glial identity is promot-ed by Notch1 signaling in the murine forebrain. Neuron 26:395–404.

Grove EA,Williams BP, Li DQ, HajihosseiniM, Friedrich A, PriceJ. 1993.Multiple restricted lineages in the embryonic rat cere-bral cortex. Development 117: 553–561.

Huang P, Xiong F, Megason SG, Schier AF. 2012. Attenuation ofNotch and hedgehog signaling is required for fate specificationin the spinal cord. PLoS Genet 8: e1002762.

Isshiki T, Pearson B, Holbrook S, Doe CQ. 2001.Drosophila neu-roblasts sequentially express transcription factors which spec-ify the temporal identity of their neuronal progeny. Cell 106:511–521.






Kalyani A, Hobson K, Rao MS. 1997. Neuroepithelial stem cellsfrom the embryonic spinal cord: isolation, characterization,and clonal analysis. Dev Biol 186: 202–223.

Kang P, Lee HK, Glasgow SM, FinleyM, Donti T, Gaber ZB, Gra-hamBH, Foster AE,Novitch BG,Gronostajski RM, et al. 2012.Sox9 and NFIA coordinate a transcriptional regulatory cas-cade during the initiation of gliogenesis. Neuron 74: 79–94.

KaoC-F, YuH-H,HeY, Kao J-C, Lee T. 2012.Hierarchical deploy-ment of factors regulating temporal fate in a diverse neuronallineage of the Drosophila central brain. Neuron 73: 677–684.

Kicheva A, Bollenbach T, Ribeiro A, Valle HP, Lovell-Badge R,Episkopou V, Briscoe J. 2014. Coordination of progenitor spec-ification and growth in mouse and chick spinal cord. Science24: 1411.

Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF.1995. Stages of embryonic development of the zebrafish.Dev Dyn 203: 253–310.

Kong JH, Yang L, Dessaud E, Chuang K, Moore DM, Rohatgi R,Briscoe J, Novitch BG. 2015. Notch activity modulates the re-sponsiveness of neural progenitors to Sonic hedgehog signal-ing. Dev Cell 33: 373–387.

Leber SM, Sanes JR. 1995. Migratory paths of neurons and glia inthe embryonic chick spinal cord. J Neurosci 15: 1236–1248.

Leber M, Breedlove SM, Sanes J. 1990. Lineage, arrangement, anddeath of clonally related motoneurons in chick spinal cord. JNeurosci 10: 2451–2462.

Li H, Paes de Faria J, Andrew P, Nitarska J, RichardsonWD. 2011.Phosphorylation regulates OLIG2 cofactor choice and themo-tor neuron-oligodendrocyte fate switch. Neuron 69: 918–929.

Liu Y, Rao M. 2003. Oligodendrocytes, GRPs and MNOPs.Trends Neurosci 26: 410–412.

Lu QR, Sun T, Zhu Z, Ma N, Garcia M, Stiles CD, Rowitch DH.2002. Common developmental requirement for Olig functionindicates a motor neuron/oligodendrocyte connection. Cell109: 75–86.

Luskin MB, Pearlman AL, Sanes JR. 1988. Cell lineage in the ce-rebral cortex of the mouse studied in vivo and in vitro with arecombinant retrovirus. Neuron 1: 635–647.

LuskinMB, Parnavelas JG, Barfield JA. 1993.Neurons, astrocytes,and oligodendrocytes of the rat cerebral cortex originate fromseparate progenitor cells: an ultrastructural analysis of clonal-ly related cells. J Neurosci 13: 1730–1750.

McCarthy M, Turnbull DH, Walsh CA, Fishell G. 2001. Telence-phalic neural progenitors appear to be restricted to regionaland glial fates before the onset of neurogenesis. J Neurosci21: 6772–6781.

Mukouyama Y, Deneen B, Lukaszewicz A, Novitch BG, Wich-terle H, Jessell TM, AndersonDJ. 2006. Olig2+ neuroepithelialmotoneuron progenitors are not multipotent stem cells invivo. Proc Natl Acad Sci 103: 1551–1556.

Noll E, Miller RH. 1993. Oligodendrocyte precursors originate atthe ventral ventricular zone dorsal to the ventral midline re-gion in the embryonic rat spinal cord. Development 118:563–573.

Orentas DM, Hayes JE, Dyer KL, Miller RH. 1999. Sonic hedge-hog signaling is required during the appearance of spinalcord oligodendrocyte precursors. Development 126:2419–2429.

Pan YA, Freundlich T, Weissman TA, Schoppik D, Wang XC,Zimmerman S, Ciruna B, Sanes JR, Lichtman JW, Schier AF.2013. Zebrabow: multispectral cell labeling for cell tracingand lineage analysis in zebrafish. Development 140:2835–2846.

Park H-C, Appel B. 2003. Delta–Notch signaling regulates oligo-dendrocyte specification. Development 130: 3747–3755.

Park H-C, Mehta A, Richardson JS, Appel B. 2002. Olig2 is re-quired for zebrafish primary motor neuron and oligodendro-cyte development. Dev Biol 248: 356–368.

Park H-C, Shin J, Appel B. 2004. Spatial and temporal regulationof ventral spinal cord precursor specification by Hedgehog sig-naling. Development 131: 5959–5969.

PringleNP, RichardsonWD. 1993.A singularity of PDGF α-recep-tor expression in the dorsoventral axis of the neural tube maydefine the origin of the oligodendrocyte lineage.Development117: 525–533.

Pringle NP, Yu WP, Guthrie S, Roelink H, Lumsden A, PetersonAC, Richardson WD. 1996. Determination of neuroepithelialcell fate: induction of the oligodendrocyte lineage by ventralmidline cells and sonic hedgehog. Dev Biol 177: 30–42.

Qian X, Goderie SK, Shen Q, Stern JH, Temple S. 1998. Intrinsicprograms of patterned cell lineages in isolated vertebrate CNSventricular zone cells. Development 125: 3143–3152.

Qian X, Shen Q, Goderie SK, He W, Capela A, Davis AA, TempleS. 2000. Timing of CNS cell generation: a programmed se-quence of neuron and glial cell production from isolated mu-rine cortical stem cells. Neuron 28: 69–80.

Reynolds BA, Tetzlaff W, Weiss S. 1992. A multipotent EGF-re-sponsive striatal embryonic progenitor cell produces neuronsand astrocytes. J Neurosci 12: 4565–4574.

Rowitch DH. 2004. Glial specification in the vertebrate neuraltube. Nat Rev Neurosci 5: 409–419.

Satou C, Kimura Y, Higashijima S-i. 2012. Generation ofmultipleclasses of V0 neurons in zebrafish spinal cord: progenitor het-erogeneity and temporal control of neuronal diversity. J Neu-rosci 32: 1771–1783.

Schmid A, Chiba A, Doe CQ. 1999. Clonal analysis ofDrosophilaembryonic neuroblasts: neural cell types, axon projectionsand muscle targets. Development 126: 4653–4689.

Shin J, Park HC, Topczewska JM, Madwsley DJ, Appel B. 2003.Neural cell fate analysis in zebrafish using olig2 BAC trans-genics. Methods Cell Sci 25: 7–14.

Simpson P. 1997. Notch signalling in development: on equiva-lence groups and asymmetric developmental potential. CurrOpin Genet Dev 7: 537–542.

Stasiulewicz M, Gray S, Mastromina I, Silva JC, Bjorklund M,Seymour PA, Booth D, Thompson C, Green R, Hall EA,et al. 2015. A conserved role for Notch in priming the cellularresponse to Shh through ciliary localisation of the key Shhtransducer, Smo. Development 142: 2291–2303.

Stolt CC. 2003. The Sox9 transcription factor determines glialfate choice in the developing spinal cord. Genes Dev 17:1677–1689.

Takebayashi H, Nabeshima YI, Yoshida S, Chisaka O, Ikenaka K,Nabeshima YI. 2002. The basic helix–loop–helix factor Olig2is essential for the development of motoneuron and oligoden-drocyte lineages. Curr Biol 12: 1157–1163.

Tawk M, Araya C, Lyons DA, Reugels AM, Girdler GC, BayleyPR, Hyde DR, Tada M, Clarke JDW. 2007. A mirror-symmet-ric cell division that orchestrates neuroepithelial morphogen-esis. Nature 446: 797–800.

Temple S. 1989. Division and differentiation of isolated CNSblast cells in microculture. Nature 340: 471–473.

Timsit S, Martinez S, Allinquant B, Peyron F, Puelles L, Zalc B.1995. Oligodendrocytes originate in a restricted zone of theembryonic ventral neural tube defined by DM-20 mRNA ex-pression. J Neurosci 15: 1012–1024.

Vallstedt A, Klos JM, Ericson J. 2005. Multiple dorsoventral ori-gins of oligodendrocyte generation in the spinal cord and hind-brain. Neuron 45: 55–67.

Ravanelli and Appel





Warf BC, Fok-Seang J, Miller RH. 1991. Evidence for the ventralorigin of oligodendrocyte precursors in the rat spinal cord. JNeurosci 11: 2477–2488.

Wu S, Wu Y, Capecchi MR. 2006. Motoneurons and oligodendro-cytes are sequentially generated from neural stem cells but donot appear to share common lineage-restricted progenitors invivo. Development 133: 581–590.

YuWP,Collarini EJ, PringleNP, RichardsonWD. 1994. Embryon-ic expression of myelin genes: evidence for a focal source of ol-igodendrocyte precursors in the ventricular zone of the neuroltube. Neuron 12: 1353–1362.

Zannino DA, Appel B. 2009. Olig2+ precursors produce abducensmotor neurons and oligodendrocytes in the zebrafish hind-brain. J Neurosci 29: 2322–2333.

ZhouQ, ChoiG, AndersonDJ. 2001. The bHLH transcription fac-tor Olig2 promotes oligodendrocyte differentiation in collabo-ration with Nkx2.2. Neuron 31: 791–807.

Zhou Q, Anderson DJ, Choi G, Anderson DJ. 2002. The bHLHtranscription factors OLIG2 and OLIG1 couple neuronal andglial subtype specification cell fate transformations in Olig1/2 mutant mice results double-mutant mice results in loss ofmotoneurons. Neuron 109: 61–73.






10.1101/gad.271312.115Access the most recent version at doi: originally published online November 19, 201529:2015, Genes Dev.

Andrew M. Ravanelli and Bruce Appel by progenitor recruitmentMotor neurons and oligodendrocytes arise from distinct cell lineages

Material

Supplemental

http://genesdev.cshlp.org/content/suppl/2015/11/18/gad.271312.115.DC1

References

http://genesdev.cshlp.org/content/29/23/2504.full.html#ref-list-1

This article cites 60 articles, 26 of which can be accessed free at:

License

Commons Creative

.http://creativecommons.org/licenses/by-nc/4.0/at Creative Commons License (Attribution-NonCommercial 4.0 International), as described

). After six months, it is available under ahttp://genesdev.cshlp.org/site/misc/terms.xhtmlsix months after the full-issue publication date (see This article is distributed exclusively by Cold Spring Harbor Laboratory Press for the first

ServiceEmail Alerting

click here.right corner of the article or

Receive free email alerts when new articles cite this article - sign up in the box at the top

© 2015 Ravanelli and Appel; Published by Cold Spring Harbor Laboratory Press


http://genesdev.cshlp.org/lookup/doi/10.1101/gad.271312.115

http://genesdev.cshlp.org/content/suppl/2015/11/18/gad.271312.115.DC1

http://genesdev.cshlp.org/content/29/23/2504.full.html#ref-list-1



http://genesdev.cshlp.org/cgi/alerts/ctalert?alertType=citedby&addAlert=cited_by&saveAlert=no&cited_by_criteria_resid=protocols;10.1101/gad.271312.115&return_type=article&return_url=http://genesdev.cshlp.org/content/10.1101/gad.271312.115.full.pdf



Documents

Motor neurons and oligodendrocytes arise from distinct ...genesdev.cshlp.org/content/29/23/2504.full.pdf · Motor neurons and oligodendrocytes arise from distinct cell lineages by